#svaixdthesis

The voter-registration deadline for the April 19 primary closed 25 days beforehand, when no candidate had even campaigned in New York, and independent or unaffiliated voters had to change their party registrations by October 9, 193 days before April 19, to vote in the closed Democratic or Republican primaries. This will disenfranchise nearly 30 percent of New Yorkers. [...] 

Version 7 of the design prototype is up!

You can view the Invision prototype live here. 

Okay now on to user testing feedback! :)

Design Feedback / User Testing

Homepage CTA (Information input flow)

“Why not just always have the user enter in their name, dob, and zip, if this will guarantee more accurate info? This information isn’t that personal.” (1/3 users)

“Before I enter in my personal info, I want to know that you won’t share it with anyone else.. so maybe include a message about that, or a link to the privacy policy?” (1/3 users)

“I’m not sure about my address.” should say “I’m not sure if or where I’m registered to vote."

In response to the first question above… There are 2 things I considered for this information flow:

I want to make it easy for power users (i.e. users who know for sure that they’re registered and where) to get straight to the info, without having to give me identifying information.

Some people will balk at being asked up front for identifying info.

In general, it seems to be the case that there are 2 types of users:

The Comfortable Voter: Slightly older — maybe mid-late 30s. Comfortable with voting. Generally knows if they’re registered or not, and where. More likely to be resistant to giving personal information. Reads more carefully.

The New Voter: Slightly younger — early to late 20s. They generally feel like they don’t know much about voting or elections. Tends to skim more than reading every word. More likely to be unsure about where they’re registered to vote or if at all. Doesn’t mind being asked to give basic personal info such as name / dob / email, etc. They are used to this from the plethora of other apps that do this.

I’m currently trying to appeal to both types of users. Perhaps with a slightly stronger leaning towards the latter, while still trying to make sure that the former doesn’t feel alienated.

Homepage Navigation

“It looks like each section of the homepage almost fills the screen on mobile… Should the edges of each section snap to the screen edges to center content?"

"Consider indicator dots to let me know how much more info is on the homepage. I wouldn’t have made it past first CTA on the homepage if this wasn’t a user test, so I wouldn’t have seen that there are other upcoming elections."

The “Send me reminders!” feature

Text VS Email — It was pretty split. Younger voters tend to prefer text, because they actually read it. Older voters tend to prefer email, because it feels less intrusive.

“How often will these be sent? I don’t want them if they’re going to be annoying, like in other apps."

For texts — Include “msg rates apply” text (1/3 users)

“You’re all set!” —> Include a note to the user that they can expect a text/email confirmation shortly.

“What are registration deadlines? Aren’t I already registered? I probably wouldn’t check this box.” (1/3 users)

“I want to see more information…"

Voter registration — “What would an inactive voter status be…?” (1/3 users)

Voter registration — “Is this actually what I would see on my official voter registration card?” (1/3 users)

Ballot — “Why do I care who a delegate is pledged to?” (1/3 users) I HAVE THIS SAME QUESTION FSAJDOLSANLDSA. And I still haven’t figured out how to answer it… ;_;

FAQ —  Add something in about absentee ballots for people who won’t be in town on the day of an election (1/3 users)

Feature requests:

Next Election — Include an “Add to calendar” feature (3/3 users) (this could be next to the date for this election… and/or it can also be part of the text/email reminders…)

View Ballot — “Even if I’m registered with the Democratic party, I still kind of want to see the ballot for the other party..” (2/3 users)

Visual design

“The header at the top doesn’t look like I can click on it.” (1/3 users)

“I thought the arrow on the homepage was a link that would scroll me down to the next area.” (1/3 users)

"The address input icon looks like an email icon. At first I thought you wanted my email address.. not my home address.” (1/3 users)

“This seems like a pro-Bernie app, because Sanders delegates are at the top of the list.. How does the order of this list get determined?” (on the page where you view the ballot) (1/3 users)

“I can’t vote in this app, right? Maybe make that more clear.” (1/3 users)

“Candidates are far down the page where i can see my info… What if I ONLY wanted to see the ballot? Although I can see why it’s like that if you intentionally want the user to see all this other information first…” (1/3 users)

“Why is the FAQ after the reminders CTA? Wouldn’t you want to answer these questions while the user is looking at their ballot?” (1/3 users)

Things I Need to Design

The other election types

When a user checks their voter registration, but they aren’t registered with that info (try a different address?)

Annnnd some positive feedback:

“Wow! The flow feels like it makes a lot more sense, and I love the language! I feel like you’re teaching people a lot about the process, without them even having to know what questions to ask, it feels really natural.” - Mat 

LA County was able to redesign and implement a new voting process because they were able to use their own funding. Should the federal government by providing funding for this sort of thing to all election jurisdictions?

First some history… historically, elections have been highly localized, and that was by design.

After the 2000s, with bush v gore, they started to nationalize more, which meant a bit more federal funding. the focus there was to replace punch card systems, because people felt there were problems with punch card systems. so this money was distributed to the states, who then distributed it to the counties.

This culminated in the Help America Vote Act of 2002.

looking back… all of that money was spent on voting systems that were not well-designed and were focused on elections administration, and a very small market place of vendors — without much regard to the actual voting experience.

fast forward to today.. here we are in 2016.. and we’re still talking about the need to modernize our elections process.

a lot of those systems ended up being scrapped… but then the economic downturn happened.. and there wasn’t any money to go around. There’s not a clear funding source. Local govts have to do what they can.

But the problem remained for the team in LA county. Their voting process really wasn’t working, and they really needed to modernize. In the end, they didn’t use much of that federal funding.. partly intentionally and partly by circumstance. There wasn’t anything on the market that they could really use. They adapted their voting system in anticipation that there would be a solution on the market.. but that never happened (see aforementioned notes). THAT’s when they decided to design it themselves.

They intentionally approached the project not as a tech or machine project.. but with the approach of “what is the ideal voting experience.” They were willing to do something radically different.

They acknowledged up front that most election jurisdictions around the country won’t have the resources / funding to do what LA county can do. LA county has wayyyy more voters than more other jurisdictions.

The team in LA county hoped that their work would be transferable.. not in terms of actual machines or anything, but in terms of the research and ideas.

So what’s the current status in LA county?

They are at the end of the design and specification phase, they’re about to begin looking for manufacturers later this year. The hope is that what they’re doing will serve to expand the markets by giving new vendors something new to work with.

One key goal that they have is to make sure that the software that goes with the voting machines will be open-source.They’re trying to use off-the-shelf components when possible too.

Do you think the federal govt SHOULD have a role?

Yes, they should play a role. They have a responsibility to set standards. The Elections Assistance Commission  would do this. There’s some debate about whether or not that should be state or federal, but I do think the federal govt should have a role. Like the recent debacle in Arizona.. the federal govt should do something there.

The other thing though is that elections should reflect their local culture. Jurisdictions are different. LA county has language issues that other counties don’t have. So there needs to be that local culture.. but there should still be standards that make the process consistent across the nation, which means the federal govt should play a larger role.

The funding question is a huge question. While state / federal govts have been more and more involved, they haven’t provided funding. Most funding is local.

Ironically, you tend to get more funding when you’re doing a bad job, then a good job. Crises generate more money… e.g. Arizona is probably in a great position to ask for money right now.

Data is another aspect made difficult by the hyper-localization of our elections system. What are your thoughts on this, and what does LA county do right now in terms of election data?

“Yeah, you’ve hit the nail on the head.”

Historically CA hasn’t been very good about aggregating statewide data. Elections have always been administered by the 58 individual counties, and the data has been separated. They are, right now, definitely in the process of trying to aggregate this data.

Who I spoke with: Bobby Hoffman + Amy Cohen

Notes:

They work directly with state election offices to try to collect polling site + ballot information.

At a bare minimum they get the poll site from the state election office.

Their goal is to ALSO get ballot info. Which is much harder.

The Ballot info is hard because where it comes from varies:

Virginia - every single election, even in small towns, is run by the state. The counties push the info up to the state.

In other states, the counties aren’t required to push this data up to the state.

In some states, even cities within a county are separate from the county and state — e.g. LA county and Texas county (LA county probably cuz it’s huge?)

So each state varies from having the data pretty centralized (e.g. Virginia) to very decentralized (e.g. California).

How do they get the states to work with them?

The provide FREE technical assistance: they help the state boards of election create tech infrastructure to help connect them with their counties and cities.

I explained to them how NationBuilder goes about collecting voter file data jurisdiction by jurisdiction. Their response:

There are potentially as many as 10,000 election jurisdictions, depending on how you count them. Doing it one by one is NOT sustainable.

The VIP method of just going to each state, and then helping states create infrastructure to connect with their cities/counties, is more sustainable.

NYC is really decentralized:

Campaign finance board manages the candidates on the ballot exclusively

Street data from office of city planning

The actual election is run by the nyc board of elections

Matching a home address to an election district is the hardest part. Office of city planning provides the street data to do this. Why? older technologies and the size of the street files (they’re super large files)

NY State Board of Election:

NY actually doesn’t need tech assistance in communicating with their cities/counties.

…so VIP just gets the info from them

VIP said that at the moment lots of existing tools are using the Google Civic Info API for poll site:

e.g. Hillary Clinton https://www.hillaryclinton.com/commit/locate

e.g. Bernie Sanders? (couldn’t find it)

(too bad it throws an error like most of the time…)

The Voter File: There’s no national voter file. Third-party companies basically spend a lot of manpower setting up an infrastructure to collect this data, and they then sell it to campaigns. This is how the do it:

Every state has their own voter file. At the top of the chain is the Secretary of State. Then next it’s the County Registrar.

The closer you can get to the source, the better the quality of the data.

The quality of the data varies greatly from state to state, largely because each state collects and maintains their data in a different way, which is largely due to the history and culture of that state and its politics.

e.g. Ohio and Minnesota are great. The data is refreshed every 2 weeks, and it’s cheap (like $1.25).

e.g. Arizona sucks. It’s like $30,000 for each time you request data from them.

Interestingly, there’s also laws about how you can use the voter file data. In most states, it’s illegal to use it for business purposes (i.e. targeting consumers). However, this is legal in Arizona, since Arizona is a very pro-business state.

The biggest difficulty is usually finding that one person who can actually give you the data

In some states, this is easy. You just call the county registrar, they tell you the price, and you buy it.

In other states, finding the right person is incredibly difficult. It took NationBuilder a month to find the right person in the state of Indiana.

In California, you have to go to each of the 54 individual counties to get this data.

RunForOffice.org: NationBuilder has a product called runforoffice.org. The way they get their data for this tool is basically a bunch of interns do a lot of phone calling.

Suggestions:

Talk to political science professors (specifically about why the upcoming primary ballot includes voting for president AND the DNC delegates)

If you can’t do a voter registration check through an API, try seeing if you can do it via a voter file from one of the smaller political parties:

Green Party is your best bet (http://www.gpny.org/)

Working Families party is worth a shot too (http://workingfamilies.org/)

Basically, these parties are small enough that they’re not terribly busy, and also small enough that they’ll probably empathize with your data problem — since they had these exact problems themselves at some point.

If you have a voter file, you should be able to search for a registered voter’s address based on their first name + last name + zip code + date of birth. 

You could also call NationBuilder to see if there’s an API for the voter file in NYC. Tell them you’re an app developer, and you can ask for the app team. They have a similar offering called the Election Center. No API has been publicly announced, but they’ll be able to give you a sense of whether or not they’ll be launching one, and if so, when. (http://nationbuilder.com/voterfile)

Maybe also check out the National Democractic Municipal Officials? (http://www.nationaldmo.org/)

Cool resources to check out:

https://www.countable.us/  |  Started in California. They explain issues and props in simple language. (because in CA they have “direct democracy” where voters vote directly on issues)

API for layman’s terms explanations of political offices?

Who works on the Google Civic Information API, specifically the election stuff? For example, I know that NYC has an ElectionAPI that provides poll site and ballot info, but this is not available in the Civic Info API. Do you know why?

I don’t work specifically on that. The team at the Voting Information Project does. Email them and ask!

How is the Google Civics organized?

Voter assistance - e.g. try googling “campaign finance” + a candidate’s name

Election night - e.g. try googling “presidential primaries"

Structured Data - This is dictated by what users search for the most on Google. e.g. if people in India often search for “felonies” then we’ll create a Google search result card with the info they’re looking for. This data is usually from a 3rd party source, we just give it structure.

How do you decide what products to build?

After a big election sprint, when we finally have a moment to breathe, we’ll usually do a 2-3 day workshop  on idea generation with the whole Civics team. We fly everyone in (some people are in different cities/countries) to NYC.

It’s usually based on what people search for, as well as what we think Google is most likely to be able to have a unique impact on. (e.g. Products that we think other people could do, we might avoid.)

Then there’s an effort estimation process, followed by resource allocation. And of course, we’re always short on time and budget.

The user we most care about is the “Interested Bystander."

Helpful Links:

Understanding America’s “Interested Bystander”: A Complicated Relationship with Civic Duty  |  Link

An article published by the Google team on the “interested bystander” user persona, based on a lot of research. This user persona drives a lot of product design decisions.

Key points:

Prior personal experience or expertise

Clear interests at stake

Seeking emotional fulfillment

Too busy, no time

Don’t know what to do

I won’t have an impact

Full report HERE Deck of key takeaways HERE

The Three Levers of Civic Engagement  |  Link

An article that Anthea wrote about why motivates people to be civically engaged.

There are 3 ways we can increase user engagement:

LOWER the COST

INCREASE the (sense of) DUTY - e.g.

“One big issue I came across in my political work in LA was people showing up to the wrong poll site. Do people know they can still vote, even if they’re at the wrong poll site?"

For data sources, try looking at the following… 

PDI - http://politicaldata.com/

Nationbuilder - http://nationbuilder.com/

NGPVAN - https://www.ngpvan.com/

“Try looking at specialty groups — who have an incentive to help voters understand the voting process — for helpful ways of organizing voter information, e.g. descriptions of offices…”

League of Conservation Voters

On how feasible it is to create a nationwide tool, and on how it might need to be done:

Google Maps is a great example. Google first started with the Maps app. Once everyone started using it, they had a bit of leverage. So they then took it to Portland, OR — their goal was to partner with the transit companies in Portland to integrate their transit data into the Google maps app.

They succeeded. But only because they had a way to incentivize the Portland transit companies. Once Portland joined, then other transit companies in cities all over started joining, and from there they just had to take it on one by one.

Point being: You have to incentivize organizations to provide you with the data you’re looking for and work with you to maintain it.

On the power of money:

Aaron thinks Google is might be having a harder time organizing voting data because there isn’t nearly as money in it as there is with transit.

Voting data is a really complex set of data

: There are lots of rules that are different depending on state, city, municipality, etc. In that way, voting data is much harder than transit data, which is probably a little easier to organize and standardize.